A.M. Oberbauer, PhD
Department of Animal Science, University of California-Davis, Davis, California
Overview of the Issue
The basic objective of all breeders is to improve on a breed, thus the axiom of "breed the best to the best." The last part of the adage, "and hope for the best" oft quoted by thoroughbred breeding legend C.V. Whitney, sums up the historical breeding perspective. Until genetic tests exist allowing breeders to identify, and then select for, particular traits, physical traits and disorders must be dealt with from a probability/likelihood viewpoint rather than a breeding certainty. While Mendelian inherited traits are more easily tracked, and possibly dealt with from a breeder's point of view, disorders that are genetically complex in their regulation often are of most concern.
Selective breeding either for or against a trait or disorder requires that the trait/disorder be under genetic control. Generally the first evidence that a trait has a genetic component is empirical: a disorder having greater prevalence among related individuals leads breeders to consider that the disorder is familial and controlled in part by genetic contributions. With some disorders, affected dogs appear in to be segregating through the generations. Alternatively, the disorder may appear sporadically in the pedigree. Familial data can be analyzed statistically to estimate the proportion of genetic vs. environmental contribution to the phenotypic expression of a trait. This heritability value provides an estimate of the extent of genetic control over a trait. Information from many different families representing multiple geographical locations yields the most accurate heritability estimate because a range of environmental influences are then evaluated. Further, the magnitude of the heritability estimate can be predictive of a breeder's ability to effectively select against the disorder.
Although a high heritability estimate has been associated with Mendelian inheritance, as more genetic studies are completed that association has become less robust; that is especially true if environmental factors are shared amongst dogs affected with a particular disorder. To assess mode of inheritance an additional statistical analysis must be done. Complex segregation analysis models whether a trait is best described as governed by a single gene inherited in a Mendelian fashion or if the disorder is more polygenic with multiple genes contributing to its expression. If it is the latter, the disorder is more difficult to eradicate due to the seemingly sporadic nature of the disorder's expression. Thus, knowledge of the mode of inheritance is critical to the success of selective breeding away from a particular genetic disorder.
In complex segregation analysis, the phenotypic data representing expression of the disorder are fit to proposed models of inheritance (single gene, few genes, many genes, many genes with a single major gene, etc.) and then the how well the data "fit" is compared among the models. That is, if a particular mode of inheritance is specified, will that model generate the observed frequency of the disorder that is seen in the actual data? The different models are then statistically compared in a maximum likelihood analysis which permits the investigator to determine what mode of inheritance is most consistent with the actual recorded data. These analyses also account for environmental factors that may influence the expression of the disorder.
Genes regulating disorders inherited in a classical Mendelian fashion can be identified with linkage analysis as discussed in other presentations at this conference. While the identification and characterization of Mendelian genetic disorders are by no means simple, similar characterizations for complex disorders present a more complicated challenge to identify the underlying genes. That is due in part, as noted above, to the lack of a clear pattern to the transmission of the disorder through the generations and to the interaction of the genes with particular environmental exposures. This ambiguity in the transmission of the disorder poses problems for predicting the risk of an individual to either contract or pass on the disorder.
Complex disorders regulated by a few genes are sometimes referred to as "oligogenic" and those regulated by a large number of genes interacting with environmental influences are referred to as "polygenic" or "multifactorial". Complex disorders are also often referred to using Mendelian terminology with additional terms such as "autosomal recessive with incomplete penetrance" or "autosomal dominant with modifiers." The variability in terminology merely underscores the multifaceted interaction of the genetic and environmental contribution to the expression of a complex disorder. In all cases, regardless of terminology, more than one gene regulates the expression of the disorder and to best select against a particular disorder requires greater knowledge than for simple disorders.
Human complex genetic disorders recently have received a great deal of attention, partly due to the relative straightforward approach necessary to identify the gene underlying a Mendelian disease. In fact, a recent review states "Nearly every Mendelian genetic disorder has now been mapped to a specific gene or set of genes" (Mayeaux, 2005). With the conquest of simple disorders, the challenge of complex disorders coupled with the availability of new genetic tools has appealed to scientists. Further, most disorders appear to be complexly inherited. This has led to more comprehensive statistical approaches and improved technology being recruited to detect the chromosomal regions behind complex disorders.
Information derived from the complex segregation analysis can guide the characterization of the genetic regulation. For example, as mentioned, one potential mode of inheritance model that can be tested is many genes with a single major gene. Further model modifications can include whether that single major gene is inherited in a Mendelian fashion. A major gene is defined as one whose contribution has a significantly large influence on the expression of a trait. (Of note, other definitions more narrowly define major gene as one that is necessary and sufficient to result in expression of the disorder.)
The presence of a major gene can facilitate the investigation and localization of the genetic components behind the disorder. The technical approach for a disorder in which a major gene exerts a large effect on expression is similar to that for a simple Mendelian disorder. Complications to the detection of the chromosomal region behind a disorder arise in the accuracy of the diagnosis of the disorder in individuals and, as plagues all genetic studies, if the disorder is late onset, the appropriate censoring of the data (e.g., is the five year old unaffected or merely "preaffected" and not yet expressing the disorder?). Disorders regulated by a major gene offer a great probability of success in actually identifying the single causal gene the significantly influences the expression of the disorder. The mutation within that gene can then be used to develop diagnostic breeding tools to aid in selection of superior broodstock.
In contrast, disorders that are polygenic with all involved genes contributing somewhat equivalently to the expression of the disorder have been, historically, extremely problematic in terms of identifying and developing DNA based selection tools for those genes. One reason for this is that naturally occurring polymorphisms in a gene may be indistinguishable from mutations. When many different genes are known to contribute to the expression, but the identity of the genes is unknown, then attempting to correlate DNA changes with the expression of the disorder is particularly difficult. In an attempt to circumvent this hurdle, scientists have created inbred lines of animals (e.g., cattle) to minimize the inherent genetic polymorphism and accentuate the genetic change that results in particular traits of interest. These studies are often referred to as quantitative trait loci (QTL) studies due to the generalized quantitative nature of complex traits.
One application of studying hereditary health disorders in dogs is in advancing our understanding of human conditions. But in most cases, advances to dog and cat genetic studies rely upon findings in mouse, livestock, or human studies. For instance, current emphasis is being placed on identifying the genes regulating human diabetes, obesity, and autism (Mayeux, 2005; Veenstra-VanderWeele et al., 2004). Thus, we can look to success in these species as indicators of successful characterization and genetic marker development for complex traits. For example, in livestock, identifying the DNA regulating marketable traits is desirable. Six separate chromosomes have been identified as regulating the thickness of fat on market cattle (Li et al., 2004). However, it is important to note that few QTL study results have been successfully implemented into mammalian breeding schemes. There remain, however, many lessons to be learned from current, and past, studies of complex traits in these species.
Although current approaches focus on identification of the underlying DNA with the objective of developing DNA based tests, the development of DNA based tests for fully polygenic complex disorders will take a great deal of time and resources. Until such genetic tests become available, breeders need to work to minimize the frequency of complex disorders in their breeds. Once a reliable estimate of heritability and then mode of inheritance is established, breeders need to act upon that information prior to the establishment of DNA tests and marker assisted selection.
The Orthopedic Foundation of America (OFA) and Canine Eye Registry Foundation (CERF) are both voluntary registries to assist breeders in selecting breeding stock that are free of phenotypic dysplasia or eye abnormalities. These data, while useful, are limited in that a quantitative value of risk associated with using a particular dog in a breeding program is not delineated. Further, the owners must assent to publishing information from their dogs and not all of an owner's dogs are assessed thereby limiting the utility of the available data.
Yet minimizing the expression of typical complex disorders, such as deafness, hip dysplasia, behavioral abnormalities, epilepsy, and some forms of cancer, is an important goal that needs to be addressed sooner than genetic testing can be made available. Methods to cope with breeding away from complex disorders can be addressed by looking at past studies of quantitative trait improvement breeding schemes in other species. Perhaps the best documented breeding approach to improve quantitative traits (i.e., complex traits) is the National Dairy Herd Improvement Association (DHIA) whose goal, when the association was founded over 85 years ago, was to provide a voluntary record keeping system for dairy producers. As an offshoot of that, volumes of data were chronicled representing a sire's genetic contributions to milk production and reproduction parameters, all of which are complex traits. These data are analyzed and used to derive breeding values for any given sire which allows for expansive improvement in complex traits: a 12-fold increase in milk production when producers participate and utilize the DHIA program in their breeding programs (National Milk Producers Federation. 1993).
Assigning genetic merit to individual animals requires a comprehensive database for as many animals and traits as possible yielding the "so-called" depth and breadth of pedigree information that is considered invaluable to informed breeding decisions. Details on many complex traits could be compiled and then animals assigned a breeding value for each trait. The precise raw data would not be revealed but the composite breeding values would be available. Note, because dogs and cats are not bred for any single trait, an animal with a low or unfavorable breeding value for one trait/disorder may be exceptional for a different trait. Therefore, breeders could weigh the pros and cons of doing a particular breeding due to the differential weight breeders apply to various attributes. That is, particular sires complement particular dams potentially expanding the number of sires used which may avoid the "popular sire syndrome".
Applying breeding values to assign merit for different traits is being applied to the breeding of service dogs. This has been ongoing for a sufficient number of years that this breeding approach can be evaluated subjectively: a greater number of dogs are being placed with people with disabilities than previously indicating the success of the breeding program. In addition, the use of breeding values was modeled to predict the successful reduction of epilepsy incidence in the Belgian Tervuren (Famula & Oberbauer, 1998).
The future holds great promise in the availability of genetic selection tools based upon genetic mutations even for complex disorders. However, it is not prudent for breeders to wait upon the development of such genetic tests if a disorder can be minimized by informed phenotypic selection. Further, some complex disorders are regulated by so many individual genes that complete characterization of all regulating mutations is unlikely. If a disorder is known to be genetically regulated, even complex disorders can be reduced in the population by generating and using breeding merit scores. Though that requires cooperation among breeders, the benefits for a breed are immense.
1. Famula TR, Oberbauer AM. 1998. Reducing the incidence of epileptic seizures in the Belgian Tervuren through selection. Prev Vet Med. 33:251-9.
2. Li C, Basarab J, Snelling WM, Benkel B, Kneeland J, Murdoch B, Hansen C, Moore SS. 2004. Identification and fine mapping of quantitative trait loci for backfat on bovine chromosomes 2, 5, 6, 19, 21, and 23 in a commercial line of Bos taurus. J Anim Sci. 82:967-72.
3. Mayeux R. 2005. Mapping the new frontier: complex genetic disorders. J Clin Invest. 115:1404-7.
4. National Milk Producers Federation. 1993. Dairy Producer Highlights. Nat. Milk Producer Fed., Arlington.
5. Veenstra-Vanderweele J, Christian SL, Cook EH Jr. 2004. Autism as a paradigmatic complex genetic disorder. Annu Rev Genomics Hum Genet.5:379-405.